News for hackers. Stuff that matters.
Re:ECC matters by colejohnson66 on Sunday January 03, 2021 @02:38PM (#25624450)
They don’t. They just measure it differently than AMD. Intel measures at base clock, but AMD measures at sustained max clock IIRC. It’s definitely deceptive, but it’s not a lie as long as Intel tells you (which they do).
Re:ECC matters by wtallis on Sunday January 03, 2021 @04:30PM (#25625435)
Intel's TDP numbers are at best an indicator of which product segment a chip falls into. They are wildly inaccurate and unreliable indicators of power draw under any circumstance. For example, here's a "58W" TDP Celeron that can't seem to get above 20W: https://twitter.com/IanCutress/status/1345656830907789312

And on the flip side, if you're building a desktop PC with a more high-end Intel processor, you will usually have to change a lot of motherboard firmware settings to get the behavior to resemble Intel's own recommendations that their TDP numbers are supposedly based on. Without those changes, lots of consumer retail motherboards default to having most or all of the power limits effectively disabled. So out of the box, a "65W" i7-10700 and a "125W" i7-10700K will both hit 190-200W when all 8 cores/16 threads are loaded.

If a metric can in practice be off by a factor of three in either direction, it's really quite useless and should not be quantified with a scientific unit like Watts.

Re:ECC matters by ksec on Sunday January 03, 2021 @04:05PM (#25625236)
It is a lie when they change the definition of TDP without telling you first and later redefined the word to different thing once they got caught.

May be we should use a new term for it, something like iTDP.

Re:ECC matters by marcosdumay on Sunday January 03, 2021 @03:35PM (#25624974)
Well, it's a power measurement that isn't total and can't be used for design... So, it's a lie.

If they gave it some other name, it would be only misleading. Calling it TDP is a lie.